Tensorpath · upskilling people on AI · 20 days · fundamentals → the 2026 frontier
From “Attention Is All You Need” to the age of AI agents.
Tensorpath is the flagship program for training professionals and students on AI — at scale. A guided 20-day journey from the very first Transformer to today’s reasoning models, multimodal systems, and agents. You learn live with an expert trainer and a personal AI coach that hand-holds you the whole way — then earn a verifiable certificate.
How the program works
Enroll & set your goal
Tell us your background and what you want from AI. Your learning is personalised from day one.
Learn live + with your AI coach
An expert trainer presents each day live; between sessions your AI coach answers questions, gives scenarios, and tells you what to learn next.
Build a capstone & get certified
Complete the 20 days and a hands-on project to earn a certificate any organisation can verify.
The 20-day journey
Every day is a complete, story-driven lesson your trainer delivers live — and you can revisit any time.
The World Before Transformers
How a single 2017 paper rewrote the future of intelligence
Inside the Transformer
Q, K, V and the machine that pays attention to everything at once
How Transformers Learn to Speak
From raw text to fluent generation: tokens, training, and the dials behind every answer
BERT and the Encoder Revolution
How half a Transformer quietly learned to read the whole sentence at once
GPT and the Generative Line
How the other half of the Transformer learned to talk — one word at a time
Scale Is a Strategy — GPT-3 and the Scaling Laws
What happens when 'just make it bigger' becomes a serious research program
Talking to the Machine: Prompting and In-Context Learning
How a few sentences of English became the most powerful programming language on Earth
Teaching Machines to Want to Help: Alignment, RLHF, and the ChatGPT Moment
How a clumsy text-predictor became an assistant the whole world fell in love with overnight
The Open-Weight Awakening: LLaMA, Mistral, and the Models You Can Own
How a leaked file in 2023 broke the monopoly and put frontier AI on your own laptop
Make It Yours — Fine-Tuning, LoRA and PEFT
How to bend a giant model to your will without setting your budget on fire
Memory for Machines — Embeddings, Vector Databases and RAG
How to give an AI your knowledge without ever retraining it
From Chatbots to Doers — Agents, Tool Use and Function Calling
The day the model stops talking and starts acting
Beyond Text — Multimodal AI and Vision
How the Transformer learned to see, read, and talk in one mind
Machines That Imagine — Diffusion and Generative Media
How AI learned to paint, film, and sing by sculpting order out of noise
Teaching Models to Think — Reasoning, Chain-of-Thought and Test-Time Compute
The day AI learned to slow down, deliberate, and think longer before answering
Bigger, Cheaper, Longer
MoE, Long Context, and the New Architectures Racing Past Attention
Trust, Truth and Safety
Evaluation, Hallucination, and Alignment at Scale
Shipping AI
Inference, Serving, Quantization, and the Economics of Production
The Frontier Has a Pulse
Agents, MCP, Computer Use, and the Model Landscape of Mid-2026
Becoming the Teacher
Capstone, the Road Ahead, and How to Teach This Course
Start your AI journey with Tensorpath
Join the professionals and students upskilling on AI through Tensorpath.
Enroll free →